OpenAI exec quits over rushed Pentagon AI deal
OpenAI robotics lead Caitlin Kalinowski resigned after the company rushed its Pentagon contract before governance frameworks were finalized, citing concerns about surveillance without judicial oversight and autonomous weapons without human authorization. The fallout widened as 900+ Google and OpenAI employees signed a solidarity letter with Anthropic, which was blacklisted by the DoD as a "supply chain risk" for refusing to remove the same safety guardrails OpenAI's deal ultimately retained.
The real story isn't one exec's exit — it's that the "ship first, govern later" instinct is playing out on a national security stage where the stakes are lethal, and even the winner came out looking shaken.
- –Kalinowski's objection wasn't anti-military; she specifically called it a governance concern, not a values one — the deal was announced before the guardrails were defined
- –The Pentagon blacklisted Anthropic as a "supply chain risk" (a label normally reserved for Huawei) for insisting on protections against mass surveillance and autonomous weapons — the exact protections OpenAI's deal includes
- –Altman admitted OpenAI "shouldn't have rushed," a rare public acknowledgment that contract speed came at the cost of process
- –The 900-person cross-company solidarity letter is unprecedented: nearly 100 OpenAI employees signed a letter defending a direct competitor's position against their own company's leadership
- –The structural problem: defense contracts create enormous commercial pressure to compress governance timelines, but in this domain, skipping governance isn't a tech debt issue — it's a war crimes liability
DISCOVERED
26d ago
2026-03-16
PUBLISHED
33d ago
2026-03-09
RELEVANCE
AUTHOR
ML_DL_RL